The One SEO thing I dont undertand

I am in the same boat as you are with regards to the learning process. From my understanding changing content would directly affect SEO. It would also depend on what KW you changed the content to focus on. What I have found to really affect one's SEO is changing platforms, like moving from opencart to WP. It leaves you with a bunch of 404s which will lead to the user to an "oops there's nothing here" webpage. I have been cleaning up all of my broken links because I was paying someone to do my SEO and the site changed platforms 3 times. Finally I took over everything from webdesign to SEO and finally things are lining up nicely. Good luck to you.

Its ok to make change as long as you are benchmarking your changes... I had position 3 before and now its a 2 or 6.. keep it or put it back the way it was. Changing themes can have a positive / negative / or nil effect. It depends on the structure of the theme. Its really like playing russian roulette. I would say worry less with the design and spend more time focusing on content. A theme is a theme is a theme.

changing content... this one puzzles me.. don't change content ( except for maybe an added update or something ). don't delete content. Just keep adding content. 3 added articles a week is 156 possible pages by the end of a years time.

This question is not easy to answer because you are basically asking the tall tell question of, how does Google's algorithm work? (or other search engines).

Changing on-page content can significantly change your current SEO rankings or it can mildly change them, if it all. This all depends on the competitiveness of the targeted keywords and geographic search terms.

Changing themes may or may not help with rankings because Google does provide "authority" when a new website is launched or designed because it is new fresh content. This is not something that should be changed often because that could easily mess with sitemaps of your website and may initiate an algorithm penalty because you are trying to "game" the system.

H1 and H2 tags can change your rankings as well as keyword density throughout each of your webpages or blogs content (remember if you are adding keywords to a page to stay under 2% percent per keyword per page or less because you could get penalized for "keyword stuffing).

My go to is usually placing a long tail keyword in an H1 or H2 tag, once in the beginning paragraph and once in the final paragraph and that is usually it.

Meta tags and titles need targeted keywords, permalinks need targeted keywords, alt-image tags need keywords, etc.

These can all change your rankings over night. I have gotten websites ranked over night on several long-tail keywords simply by changing the on-page. The answer to your question though is, there is not definitive time frame for how long a change can increase or decrease your rankings. SEO is an ever changing process.

You will always have to change keywords and such in order to perfect your websites SEO. Once you have gotten the right targeted market down you can perform a bit more advanced SEO techniques that can spread your websites reach across national markets and increased web traffic keyword queries, not just local markets with low search queries.

SEO is a patient man/woman's game. As I tell all my clients, SEO is a marathon not a sprint.

From the various blog posts on SearchEngineWatch, SearchEngineLand, and similar sites I've noticed that constantly making changes to the backend of the site affects SEO but it rarely affects negatively given that the changes adhere to common web standards and google parameters.
By constantly, I assume monthly. However, if someone posts on a daily basis then the crawler has to go through your site everyday which might take time, however, this is arguable. What I would advise is to test run of the site to check for any markup errors or other bugs before making the website or blog live to public. This way you can ensure that impact of the changes would be positive in terms of SEO.
BTW, you have to do SEO to notice effects.